About LavishLabs

Where computer vision meets interior design.

Our Story

Built by Designers,
Powered by AI

LavishLabs was founded in 2024 in New York by Rishab Marocha after countless frustrating hours spent browsing furniture websites, wondering "will this actually fit?" and "does this match my style?"

The idea was simple: what if you could just take a photo of your room and instantly see furniture that's perfect for your space—with direct links to buy it? No more guessing dimensions. No more style mismatches. No more endless scrolling.

Today, our team of 4 engineers, designers, and AI researchers is on a mission to revolutionize how people furnish their homes. We believe everyone deserves a beautifully designed space, and technology should make that easier, not harder.

LavishLabs LLC is headquartered in New Hyde Park, New York, and we're proud to be building the future of furniture discovery.

2024 Founded
NYC Based In
4 Team Members
300+ Beta Users
Modern outdoor living space
Our Technology

The AI Behind LavishLabs

Our proprietary AI pipeline combines multiple state-of-the-art models to deliver accurate, personalized furniture recommendations.

Computer Vision & Depth Estimation

We utilize MiDaS (Mixed Data Sampling) for monocular depth estimation, allowing us to infer 3D room geometry from a single 2D photo. Combined with YOLO v8 for real-time object detection, we can identify existing furniture, architectural features, and available floor space with high accuracy.

Semantic Segmentation

Our Segment Anything Model (SAM) integration provides pixel-precise segmentation of room elements. This enables us to understand exactly where walls meet floors, identify window placements, and map out traffic flow patterns—critical for furniture placement recommendations.

Style Recognition & Embedding

We've fine-tuned a CLIP (Contrastive Language-Image Pre-training) model on millions of interior design images to create style embeddings. This allows us to understand aesthetic preferences like "mid-century modern" or "Scandinavian minimalist" and match them to our furniture database.

Product Matching Engine

Our proprietary vector similarity search powered by Pinecone indexes over 50,000 furniture products. When you upload a room photo, we generate embeddings that query this database in real-time, returning products that match your space dimensions, style, and budget constraints.

AR Visualization Pipeline

For augmented reality previews, we leverage ARKit (iOS) and ARCore (Android) combined with custom 3D model rendering. Our system automatically scales furniture models to accurate real-world dimensions using the depth data from our initial room analysis.

Recommendation Intelligence

Beyond visual matching, our transformer-based recommendation model considers factors like color harmony, proportion theory, and design principles. It's trained on datasets of professionally designed rooms to ensure recommendations aren't just matching—they're tasteful.

Get in Touch

Have questions, feedback, or partnership inquiries? We'd love to hear from you.

rishab@lavishlabs.io
516-698-0333
83 Dennis Street
New Hyde Park, NY 11040
Modern apartment interior